For smallholder rice farmers, higher farm productivity does not necessarily translate into higher household income. This paper examines this gap using evidence from the IFAD-supported Value Chain Development Programme (VCDP) among rice farmers in Niger, Taraba, and Benue States, Nigeria. Primary survey data from 282 farmers are combined with five key-informant interviews and four focus group discussions. Logistic regression is used to examine programme participation, OLS models estimate the conditional associations of programme participation and training with post-programme rice productivity, and Firth logistic models examine reported household-income improvement. Programme participation is positively associated with post-programme rice productivity, corresponding to approximately 11.7% higher predicted yield, although the association is only marginally significant. Training has a larger and statistically significant association with productivity, corresponding to approximately 35.8% higher predicted yield among trained farmers. Neither programme participation nor training is significantly associated with reported household-income improvement, and their interaction is also insignificant. The financial-constraint index is negatively associated with participation, productivity, and income improvement. The findings indicate that programme participation alone does not capture the productive differences associated with value-chain support. Training is more closely associated with productivity, while financial and production constraints remain associated with weaker outcomes. The findings also show that farm-level productivity gains cannot be assumed to translate into equivalent household-income gains.